Electrophoresis 2015, 36, 355–362
Steffen Halbgebauer1 Ute Haußmann2 Hans Klafki2 Hayrettin Tumani1 Jens Wiltfang3 ∗ Markus Otto1 ∗ 1 Department
of Neurology, University of Ulm, Germany Essen, Department of Psychiatry and Psychotherapy, Faculty of Medicine, University of Duisburg-Essen, Germany 3 Department of Psychiatry and Psychotherapy, University Medical Center (UMG), Georg-August-University, ¨ Gottingen, Germany 2 LVR-Klinikum
Received July 16, 2014 Revised September 23, 2014 Accepted October 3, 2014
Capillary isoelectric focusing immunoassay as a new nanoscale approach for the detection of oligoclonal bands The detection of oligoclonal bands (OCBs) in cerebrospinal fluid is an indicator of intrathecal synthesis of immunoglobulins which is a neurochemical sign of chronic inflammatory brain diseases. Intrathecally synthesized IgGs are typically observed in patients with multiple sclerosis. The current standard protocol for the detection of OCBs is IEF on agarose or polyacrylamide gels followed by immunoblotting or silver staining. These methods are time consuming, show substantial interlaboratory variation and cannot be used in a high throughput-approach. We have developed a new nanoscale method for the detection of OCBs based on automated capillary IEF followed by immunological detection. Evidence for intrathecal IgG synthesis was found in all tested patients (n = 27) with multiple sclerosis, even in two subjects who did not have oligoclonal bands according to standard methods. The test specificity was at 97.5% (n = 19). Our findings indicate that the novel OCB-CIEF-immunoassay is suitable for the rapid and highly sensitive detection of OCBs in clinical samples. Furthermore, the method allows for a higher sample throughput than the current standard methods. Keywords: Capillary IEF / Intrathecal IgG synthesis / Multiple sclerosis / Oligoclonal bands DOI 10.1002/elps.201400339
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction The detection of intrathecally produced immunoglobulins in cerebrospinal fluid is of utmost importance to diagnose a chronic inflammatory process in the central nervous system. In the field of autoimmune and inflammatory diseases of the central nervous system, multiple sclerosis (MS) plays a major role. MS is a multifocal demyelinating autoimmune disorder with progressive neurodegeneration and severe neuropathological hallmarks, such as demyelination and inflammatory lesions in the white matter [1–4]. The invasion of inflammatory cells across the blood brain barrier leading to oligoclonal expansion of B and T cells and ultimately degradation of myelin sheaths plays an important role in MS . In particular, clonally expanded B cells, present in the brain parenchyma  and in meningeal follicles  of Correspondence: Professor Markus Otto, Department of Neurology, University of Ulm, Oberer Eselsberg 45, 89081 Ulm, Germany E-mail: [email protected]
Abbreviations: HRP, horseradish peroxidase; MS, multiple sclerosis; OCB, oligoclonal band C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
MS patients can also be found in the CSF and are responsible for intrathecally synthesized immunoglobulins [8–14]. In almost all MS cases (⬎95%) [15, 16], intrathecally synthesized IgGs can be detected. IgGs separated by IEF form an oligoclonal band (OCB) pattern which serves as a supportive criterion in the diagnosis of MS [17, 18]. So far, the standard method for analyzing intrathecally produced OCBs is IEF on agarose/polyacrylamide gels with subsequent immunoblotting or silverstaining . However, this approach which has been used routinely since the 1970s, has its limitations. The method is highly dependent on the expertise of the lab technician and suffers from lack of automation, low throughput, variable sensitivity, and especially subjective evaluation [19,20]. However, for rapid and exact diagnosis, as an objective reference standard for round-robins tests and for multicenter trails an objective method is urgently needed. Here, we present a new approach for the detection of OCBs by automated capillary IEF with an immunological readout. The method was established as a novel application of a commercial nanoscale capillary IEF platform (NanoPro 1000) . For method establishment, we analyzed 21 ∗
These authors equally contributed as senior author.
Colour Online: See the article online to view Figs. 1–5 in colour.
S. Halbgebauer et al.
suspected chronic inflammatory disease patients regarding intrathecal production of IgGs and compared the results with the OCB patterns obtained by a standard method. In a second validation step we additionally measured the intrathecal IgG synthesis of 27 clinically well-defined MS cases and compared these results with 19 noninflammatory control patients. Our observations indicate that the measurement of OCBs by the novel CIEF-immunoassay is more objective, fast, and outmatches the standard method in terms of sample throughput and sensitivity.
2 Materials and methods 2.1 Patients All CSF and serum samples analyzed in this study were taken from patients attending the department of neurology (University of Ulm, Department of Neurology) between 2010 and 2013. Collection and analysis of CSF samples were approved by the Ethics Committee in Ulm. All individuals gave written informed consent to their participation in the study and underwent a clinical, neurological, and neuroradiological examination. Altogether we investigated 67 patients. For set-up of the protocol we investigated 21 patients with suspected chronic inflammatory diseases (mainly MS). In the validation phase we investigated 27 patients (female/male: 16/11, age mean 37.85 – SD: 10.78) with MS (diagnosis was made according to recent criteria ) and 19 patients with other neurological diseases (noninflammatory, (female/male: 10/9, age mean 41.84 – SD: 18.75)). 2.2 Materials and reagents All NanoPro 1000 reagents and materials were obtained from ProteinSimple (Santa Clara, USA). NaOH, H2 SO4 , H2 O2 , and glacial acetic acid were purchased from Merck (Darmstadt, Germany). Powdered milk was bought from Roth (Karlsruhe, Germany). Nitrocellulose and sodium acetate were obtained from Applichem (Darmstadt, Germany). 3-Amino9-ethylcarbazol was from Sigma-Aldrich (Steinheim, Germany). Normal saline was obtained from B. Braun Melsungen AG (Melsungen, Germany). Pierce Protein G Agarose was bought from Thermo Scientific. Albumin and IgG depletion Kit was purchased from GE Healthcare. Substrate solution: 26.5 mM sodium acetate-glacial acid pH 5.1, 2.38 M 3-Amino-9-ethylcarbazol (solved in ethanol), 0.4% H2 O2 . 2.3 Conventional OCB analysis by IEF on polyacrylamide gels with subsequent immunoblotting 2.3.1 Sample preparation The IgG concentrations in the CSF and the corresponding serum samples as determined by nephelometry were adjusted C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Electrophoresis 2015, 36, 355–362 Table 1. Formulas for calculation of suitable dilutions of CSFand serum samples
Up to 40 mg/L: no dilution >40 mg/L:
Always in two steps
level i n CSF( mg L ) 3
level i n serum( mg L ) 20
∗ 10 − 10
=x L 0.9% NaCl + 10 L Serum (Dil .1) =x L 0.9% NaCl + 10 L CSF
=x L 0.9% NaCl + 10 L Dil .1
by dilution with 0.9% NaCl according to the formulas listed in Table 1 (from Reiber/G¨ottingen). The final IgG concentration in the diluted CSF and serum samples was approximately 10–40 g/mL. For reasonable comparisons of CSF samples and corresponding sera, similar amounts of IgG have to be loaded . 2.3.2 Isoelectric focusing For IEF we used the multiphor II (GE Healthcare, Uppsala, Sweden). The focusing plate was cooled to 12°C and about 1 mL of petroleum was applied. On top of the petroleum a polyacrylamide gel was placed. On the anode side a strip soaked with acid sulfur (0.05 M) and on the cathode side a sodium hydroxide (1 M) buffer strip was added. Two centimeter from the anode buffer strip the loading strip was placed. Then 12.5 L diluted CSF and serum of each patient and additionally positive and negative controls were loaded on each gel. For positive and negative controls the CSF of OCB positive and negative samples was used. To start the IEF a voltage of 1200 V and 15 mA current was applied. The samples focused for 1 h and 50 min at constant charge and current. 2.3.4 Transfer onto Nitrocellulose After IEF the proteins were blotted onto a nitrocellulose membrane. For this purpose, the gel was placed on a plastic board and the membrane was put on top of it. Filters (Schleicher & Sch¨ull) were applied on top of the membrane and a 1 kg weight was deployed. The gel was blotted for 30 min. 2.3.5 Antibody incubation After blotting, the membrane was blocked in blocking solution A (2% milk in NaCl (0.9%)) for 30 min. Next, the membrane was shortly rinsed with water and then washed with 0.9% NaCl for 10 min. Again, the membrane was rinsed with water and then 50 mL blocking solution B (5 mL blocking solution A + 45 mL NaCl (0.9%)) + 30 L of an anti-human IgG antibody (Dianova, Hamburg, Germany) was applied. This equates to an antibody dilution of 1/1666. The membrane was incubated in the antibody dilution for 30 min under gentle agitation. www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2015, 36, 355–362
2.3.6 Detection The membrane was rinsed with water, washed with 0.9% NaCl for 10 min and again rinsed with water. Subsequently, the blot was incubated with substrate solution for 10–30 min until the control bands were clearly visible. After rinsing with water, the membrane was dried prior to visual inspection. The OCBs were determined by experienced employees who have evaluated the presence of OCBs over many years. This was than controlled by a physician with a certificate of the German society of neurochemistry and CSF diagnostics. In cases with discrepancies repeated analyses were performed. The final decision was then made by the physician. The 67 cases for this study were reevaluated independently by two raters (technician, SH). Eight patients had to be measured again because of discrepancies which lead to an inter-rater reliability of 88%.
2.4 OCB detection by CIEF-immunoassay After separating the IgGs in the capillaries by charge they were immobilized at their pI by exposure to UV light . Next the capillary was flowed through by a primary antibody labeled with horseradish peroxidase (HRP), forming immunocomplexes with IgGs. The capillary was washed and then, for detection, chemiluminescent reagents were flowed through the capillary reacting with HRP, they generated light which was detected by a CCD camera. Up to 96 capillaries can be processed in one run. For the analysis of OCBs by CIEF-immunoassay, the IgG concentration of each CSF and serum sample was measured by nephelometry and adjusted to approximately 10–40 g/mL by dilution with 0.9% NaCl according to the formulas in Table 1. As the CIEF approach is highly sensitive the samples were further diluted 1:10 with 0.9% NaCl. Ampholyte premix G2 (pH 3–10) was mixed with pI standards ladder 1 (1 L/44 L premix) and pI standard 5.5 (1 L/200 L premix) and vortexed thoroughly. One volume of the sample was mixed with three volumes of premix and 1 L of “DMSO inhibitor mix” was added per 50 L sample/premix mixture achieving an IgG concentration in the capillary of about 0.25 to 1.0 g/mL. Only very small sample volumes were required for this preparation. The reaction tubes were vortexed vigorously for at least 15 s. For the CIEF immunoassays, the NanoPro 1000 device (ProteinSimple, Santa Clara, USA) was employed. A total of 10 L of each premix/sample mixture was pipetted into the wells of a 384 well plate. Subsequently, polyclonal rabbit anti-human IgG HRP antibody (Dianova, Hamburg, Germany) was diluted 1:100 v/v with antibody diluent and 15 L per well were added to the 384 well-plate in a different row. Equal amounts of luminol and peroxide (Protein Simple) were mixed and 15L were loaded per well in a third row. After loading the plate was centrifuged for 10 min at 1500 g and 4°C. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
The automated assay was programed in the Compass software (ProteinSimple). The samples were loaded for 25 s, corresponding to approximately 1 L per capillary and separated by IEF for 40 min with 21 000 W. Analytes and standards were subsequently immobilized in a photochemical reaction within 100 s, followed by two washing steps (load: 20 s, soak: 150 s). The detection antibody was loaded within 2 s and incubated for 120 min. After two washing steps, luminol/peroxide was loaded for 2 s and chemiluminescence signals were detected at six different exposure times (30, 60, 120, 240, 480, and 960 s). In one run up to 96 samples or 24 CSF/serum pairs can be measured in duplicates. pI values of the analytes were calculated by the Compass software on the basis of the positions of fluorescent pH standards in each capillary. Chemiluminescence signals were automatically plotted against the pH gradient. For peak identification, a minimum S/N of 10 was required. Additionally, the approximate peak width (at full width, half max of a peak) was set to 5 (in pixels). The peaks were regarded identical if the pI deviation was not higher than 0.05 pH units. For the evaluation of the electropherograms, the data were exported to a spreadsheet application (Microsoft Excel) and the number of peaks in the CSF samples were compared with those in the corresponding sera. We classified samples to be positive for intrathecal IgG production when at least two additional peaks were found in the CSF sample compared to the corresponding serum. As an additional criterion the signal of those two peaks had to be stronger in the CSF compared to the serum and on the basic side of pI 7.7. In borderline cases where the CSF signal was at some points stronger than the serum but the software did not detect a peak, a manual follow up check was performed. In our study two raters (SH, MO) analyzed the data. Two cases were rated differently leading to an inter-rater reliability of 97.0%. All samples where measured in duplicates. In the rare cases where the duplicates lead to two different diagnoses the sample was measured again. However, this only was the case in four out of 67 overall measured samples.
3 Results 3.1 Detection of oligoclonal bands by capillary IEF immunoassay CSF samples and corresponding sera from patients suspected to have a chronic inflammatory disease were subjected to IEF in microcapillaries followed by photochemical immobilization to the inner capillary wall and immunological detection with HRP labeled polyclonal anti-human IgG antibodies. The procedure is a novel application of the commercial NanoPro1000 platform (ProteinSimple). Two examples of OCB-CIEFimmunoassay results, one from an intrathecally IgG producing patient and one from a subject negative for intrathecally synthesized IgGs are shown as electropherograms in Fig. 1 (for a display of the same data in a format mimicking a classical IEF-immunoblot, see Supporting Information Fig. 1). The www.electrophoresis-journal.com
S. Halbgebauer et al.
Electrophoresis 2015, 36, 355–362
Figure 2. The five different patterns of oligoclonal bands. IEF on polyacrylamide gels with subsequent immunoblotting. *, CSF
Figure 1. Capillary IEF electropherograms. CSF and serum of an intrathecally IgG producing patient (A) and a patient negative for intrathecal IgGs (B). In the electropherogram of the intrathecal IgG producing patient the CSF peaks show a clearly stronger signal compared to the serum and had pI values > 7.7. In the OCB negative electropherogram the found peaks were more acidic than pI 7.7 and the serum displayed stronger signals compared to the corresponding CSF. The exception is a peak around 6.5 marked with an *. This weak peak is often found in the CSF both in positive and in negative OCB patients. Black arrows indicate the ten clear IgG peaks found in the CSF. pI, isoelectric point; Blue, CSF; Green, serum. Exposure time 30 s.
electropherograms show in blue color the IgGs found in the CSF samples and in green those in the corresponding sera. We classified the patient displaying ten IgG peaks in the CSF and none in the serum (Fig. 1 top) as intrathecal IgG producing positive. In contrast, the second patient did not show extra IgG peaks in the CSF (Fig. 1 bottom) and was therefore classified as intrathecal IgG producing negative. For the discrimination we followed the international consensus criteria. According to these criteria two or more OCBs in the CSF are considered a sign for intrathecally synthesized IgGs if they are not simultaneously demonstrable in the corresponding serum . In these criteria five different prototypical banding patterns (Type 1–5) can be distinguished by classical IEF in polyacrylamide or agarose gels followed by silver staining or immunoblotting. These prototypical pattern are (1) no bands in CSF, (2) OCBs in CSF only, (3) identical OCBs in CSF and serum with additional bands in CSF, (4) identical OCB patterns in CSF and serum, (5) monoclonal bands in CSF and serum (Fig. 2). Importantly, only patterns 2 and 3 indicate an intrathecal synthesis of IgGs. Type 1, 4, and 5 C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
are in clinical work-up often simply called “oligoclonal bands negative.” In agreement with the above mentioned consensus criteria, we classified a sample positive for intrathecal IgG production when at least two additional OCBs in the CSF were found. To test for assay specificity and to further characterize the peaks detected by the CIEF immunoassay, we subjected CSF and serum from a subject positive for intrathecal IgG production to preanalytical depletion of albumin + IgG or IgG alone and compared the resulting electropherograms to those of the corresponding untreated samples. While depletion of albumin + IgG eliminated all peaks observed in the untreated samples, a very broad signal in the pI range 6 – 7 remained after IgG depletion with Protein G (Fig. 3).
3.2 Capillary IEF compared to standard methods for OCB detection The CIEF-immunoassay results obtained from the analysis of 21 clinical CSF samples and their corresponding sera were evaluated by comparing the IgG peaks in the CSF samples and corrresponding sera. On that basis, the subjects were classified as either intrathecal IgG producing positive or negative (see methods). The findings were then compared with the diagnoses obtained by IEF on polyacrylamide gels with subsequent immunoblotting (Fig. 4). All three samples in the cohort showing no OCBs according to the standard analysis (i.e. Type 1 banding pattern) were also classified negative for intrathecal IgG synthesis by the CIEF-immunoassay analysis. Due to the superior www.electrophoresis-journal.com
Electrophoresis 2015, 36, 355–362
CE and CEC
Figure 3. Characterization of the signals by preanalytical depletion of albumin + IgG or IgG alone: CSF and Serum samples from a subject postitive for intrathecal IgG production were analyzed without preanalytical treatment (panel A), after depletion of albumin and IgG (panel B) or IgG only (panel C). Blue electropherograms, CSF; Green electropherograms, serum; pI, isoelectric point. Exposure time 120 s.
detection sensitivity of our CIEF method we observed some IgG peaks in the CSF and serum samples of these OCB negative subjects while no IgG bands were detected by the standard approach. However, all of the IgG peaks found in these particular CSF samples were also present in the corresponding sera, thus classifying these samples as intrathecal IgG producing negative (Example Fig. 4 upper left corner). Three out of four samples diagnosed type 4 by classical IEF-immunoblot analysis also showed identical OCBs in the corresponding CSF samples and sera in the CIEF analysis. In the fourth sample, we observed by CIEF-immunoassay two additional OCBs in the CSF, thus classifying it as intrathecally IgG producing positive (see Supporting Information Fig. 2). All of the CSF IgG peaks detected in the electropherograms of type 1 and type 4 patients displayed pI values ⬍ 7.7 (Fig. 1 bottom, Fig. 4 upper left and lower right corner) with one exception: In the single type 4 case which we declared intrathecal producing positive, we also found IgG peaks more alkaline than pI 7.7 (see Supporting Information Fig. 2). C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Out of 14 samples diagnosed positive for intrathecal IgG production according to the standard method (Types 2 and 3), 13 were also classified as intrathecal IgG producing samples by capillary IEF. In the one remaining sample, only one additional OCB in the CSF was detected with the CIEFimmunoassay, therefore, according to the consensus criteria, classifying it as negative for intrathecal IgG synthesis. It attracted our attention that in all 14 patients the additional CSF IgG peaks were nearly all found on the basic side of 7.7 (Fig. 1 top, Fig. 4 upper right and lower left corner).
3.3 Validation of the approach with a clinical cohort In a next step we performed an analysis with a validation cohort. We included 27 clinically well-defined MS patients in our study. Two of these 27 cases were chosen because of their negative OCB results according to the analysis by IEF on polyacrylamide gels with subsequent immunoblotting. The www.electrophoresis-journal.com
S. Halbgebauer et al.
Electrophoresis 2015, 36, 355–362
Figure 4. Comparison of CIEF CSF and serum electropherograms with the corresponding immunoblot made by standard methods. One case of each OCB type (1–4) is shown exemplarily. Blue, CSF; Green, serum; pI, isoelectric point; *, CSF. Exposure time 30 s.
other 25 were diagnosed with Type 2 or Type 3 OCB-patterns. In addition 19 noninflammatory controls showing an OCB type 1 pattern by standard diagnosis were used as controls. In all cases two observations were obvious. First, in the electropherograms of all MS patients the CSF curves showed stronger signals than the corresponding sera curves. Second, in all MS cases the important additional IgG peaks in the CSF were more alkaline than pI 7.7 whereas the few CSF peaks which we could detect in controls appeared mostly on the acidic side of pI 7.7. Examples are shown in Fig. 1. Taking these findings as additional criteria (see methods) we detected by the novel CIEF approach intrathecal IgG production in 27 of the 27 MS cases, matching a sensitivity of 100%. In the control group 18 out of 19 cases showed no intrathecal IgG production (specificity 94.7%). Presumably due to the improved detection sensitivity, our CIEF method revealed C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
the presence of OCBs also in those two MS patients who had been classified intrathecally IgG producing negative by the classical standard method (Fig. 5).
4 Discussion We report here on a novel approach for the detection of oligoclonal IgG bands in CSF by capillary IEF with immunological readout. We believe that the new method offers several advantages as compared to the classical analysis by IEF on acrylamide or agarose gels followed by immunoblotting. The novel CIEF-immunoassay allows for highly sensitive and fast detection of OCBs with high sample throughput and with minimal sample volume requirement. We used CSF and serum samples of 21 patients with a suspected chronic www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2015, 36, 355–362
of ampholytes from different suppliers we tried to solve this problem. However, in our approach the reproducibility of results using a self made pH gradient with ampholytes from two different manufacturers suffered greatly which is why we cast away this idea and used the commercially available pH 3–10 gradient. All in all, we conclude that capillary IEF with an immunological readout on the NanoPro 1000 can strongly facilitate the detection of OCBs. The fully automated mode of operation allows for highly sensitive and comparably objective analysis in high-throughput settings and with minimal sample volume requirements. This might be especially relevant for multicenter trails for objective description of laboratory findings but also in clinical practice as in patients diagnosed with MS a relevant immunosuppression is performed.
Figure 5. Capillary IEF electropherograms. CSF and serum of two clinically well characterized MS cases diagnosed negative for OCBs by standard methods. In both cases two or more additional OCBs in the CSF compared to the serum were observed by CIEF; pI, isoelectric point. Exposure time 30 s.
inflammatory disease to set-up the protocol for our new approach before we assessed the diagnostic performance by analyzing a validation cohort of 27 well-defined MS patients and 19 controls. In the set-up phase with the 21 suspected chronic inflammatory disease cases 19 were diagnosed in accordance with the standard method regarding intrathecal IgG synthesis. Since the CIEF platform is very sensitive, we can assume that in the two deviating cases, oligoclonal bands were missed or falsely interpreted with the standard method. For these cases, further clinical history will be observed thoroughly. Our method can therefore be especially helpful in distinguishing difficult cases in which the presence or absence of only one or two oligoclonal bands can be decisive for the diagnosis. In the validation cohort we demonstrated that our method reliably detects intrathecal IgG synthesis in MS patients (sensitivity 100%) and is thereby even more sensitive than IEF with subsequent immunoblotting (sensitivity about 95% [15, 22, 23]). Furthermore, the specificity of 94.7% exceeds the literature value of 86–87% [15,24]. In addition, we found that two well characterized MS cases despite showing no OCBs after standard method measurement, displayed an intrathecal production of IgG which was detectable by our CIEF approach. In future experiments, the gradient of ampholytes should be improved if the small drop in the electropherograms at pH 7.2–7.4 caused by a lack of ampholytes will be addressed by the manufacturer. Using a self made gradient with a mix C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
This work was supported by BMBF (German Federal Ministry for Education and Research): Kompetenznetz Multiple Sklerose (KKMS), Kompetenznetz Neurodegenerative Demenzen (KNDD, FTLDc), and JPND projects: BIOMARKAPD and SOPHIA), PURE (Protein Research Unit Ruhr within Europe) from the State Government North Rhine-Westphalia, and the EU (NADINE) (Contract No. 246513). We especially thank our patients for participating in these studies and C. Janßen and R. Aksamija for expert technical assistance. The authors have declared no conflict of interest.
5 References  Stadelmann, C., Wegner, C., Bruck, W., Biochim. Biophys Acta 2011, 1812, 275–282.  Bruck, W., Stadelmann, C., Neurol. Sci. 2003, 24(Suppl 5), S265–267.  Kutzelnigg, A., Lucchinetti, C. F., Stadelmann, C., Bruck, W., Rauschka, H., Bergmann, M., Schmidbauer, M., Parisi, J. E., Lassmann, H., Brain 2005, 128, 2705–2712.  Bitsch, A., Schuchardt, J., Bunkowski, S., Kuhlmann, T., Bruck, W., Brain 2000, 123, 1174–1183.  Compston, A., Coles, A., Lancet 2008, 372, 1502–1517.  Ozawa, K., Suchanek, G., Breitschopf, H., Bruck, W., Budka, H., Jellinger, K., Lassmann, H., Brain 1994, 117, 1311–1322.  Serafini, B., Rosicarelli, B., Magliozzi, R., Stigliano, E., Aloisi, F., Brain Pathol. 2004, 14, 164–174.  Qin, Y., Duquette, P., Zhang, Y., Talbot, P., Poole, R., Antel, J., J. Clin. Invest. 1998, 102, 1045–1050.  Baranzini, S. E., Jeong, M. C., Butunoi, C., Murray, R. S., Bernard, C. C., Oksenberg, J. R., J. Immunol. 1999, 163, 5133–5144.  Colombo, M., Dono, M., Gazzola, P., Roncella, S., Valetto, A., Chiorazzi, N., Mancardi, G. L., Ferrarini, M., J. Immunol. 2000, 164, 2782–2789.  Monson, N. L., Brezinschek, H. P., Brezinschek, R. I., Mobley, A., Vaughan, G. K., Frohman, E. M., Racke, M. K., Lipsky, P. E., J. Neuroimmunol. 2005, 158, 170–181.
S. Halbgebauer et al.
 Obermeier, B., Mentele, R., Malotka, J., Kellermann, J., Kumpfel, T., Wekerle, H., Lottspeich, F., Hohlfeld, R., Dornmair, K., Nat. Med. 2008, 14, 688–693.  Owens, G. P., Bennett, J. L., Lassmann, H., O’Connor, K. C., Ritchie, A. M., Shearer, A., Lam, C., Yu, X., Birlea, M., DuPree, C., Williamson, R. A., Hafler, D. A., Burgoon, M. P., Gilden, D., Ann. Neurol. 2009, 65, 639–649.  Obermeier, B., Lovato, L., Mentele, R., Bruck, W., Forne, I., Imhof, A., Lottspeich, F., Turk, K. W., Willis, S. N., Wekerle, H., Hohlfeld, R., Hafler, D. A., O’Connor, K. C., Dornmair, K., J. Neuroimmunol. 2011, 233, 245–248.  Freedman, M. S., Thompson, E. J., Deisenhammer, F., Giovannoni, G., Grimsley, G., Keir, G., Ohman, S., Racke, M. K., Sharief, M., Sindic, C. J., Sellebjerg, F., Tourtellotte, W. W., Arch. Neurol. 2005, 62, 865–870.  Link, H., Kostulas, V., Clin. Chem. 1983, 29, 810–815.  Polman, C. H., Reingold, S. C., Banwell, B., Clanet, M., Cohen, J. A., Filippi, M., Fujihara, K., Havrdova, E., Hutchinson, M., Kappos, L., Lublin, F. D., Montalban, X., O’Connor, P., Sandberg-Wollheim, M., Thompson, A. J., Waubant, E., Weinshenker, B., Wolinsky, J. S., Ann. Neurol. 2011, 69, 292–302.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Electrophoresis 2015, 36, 355–362
 Tumani, H., Deisenhammer, F., Giovannoni, G., Gold, R., Hartung, H. P., Hemmer, B., Hohlfeld, R., Otto, M., Stangel, M., Wildemann, B., Zettl, U. K., Ann. Neurol. 2011, 70, 520; author reply 521.  Franciotta, D., Lolli, F., Clin. Chem. 2007, 53, 1557–1558.  Link, H., Huang, Y. M., J. Neuroimmunol. 2006, 180, 17–28.  O’Neill, R. A., Bhamidipati, A., Bi, X., Deb-Basu, D., Cahill, L., Ferrante, J., Gentalen, E., Glazer, M., Gossett, J., Hacker, K., Kirby, C., Knittle, J., Loder, R., Mastroieni, C., Maclaren, M., Mills, T., Nguyen, U., Parker, N., Rice, A., Roach, D., Suich, D., Voehringer, D., Voss, K., Yang, J., Yang, T., Vander Horn, P. B., Proc. Nat. Acad. Sci. U S A 2006, 103, 16153–16158.  McLean, B. N., Luxton, R. W., Thompson, E. J., Brain 1990, 113, 1269–1289.  Ohman, S., Ernerudh, J., Forsberg, P., Henriksson, A., von Schenck, H., Vrethem, M., Ann. Clin. Biochem. 1992, 29, 405–410.  Paolino, E., Fainardi, E., Ruppi, P., Tola, M. R., Govoni, V., Casetta, I., Monetti, V. C., Granieri, E., Carreras, M., J. Neurol. Neurosurg. Psychiatry 1996, 60, 572–575.